forecasting underlying mkt vol, from modeling of greeks 'term structure' or cumulative measure

[JTDerp]

What research topics might be good starting points to consider the feasibility of forecasting volatility in the underlying market? I know that many traders model the 3D distribution of option greeks vs expiry & moneyness/strike and apply other factors in the overall model to forecast, so this is a question of those other factors.

I've read through 2 papers which researchers at Nomura/NCI wrote with mentions of cumulative gamma by-strike, and postulate some 'if-touched' barrier/boundary scenarios where if the UL breaches a level where this cum. gamma has the highest sum, the volatility is very likely to rise as participants adjust positions. Seems similar to 'sticky-delta' like Derman and others have talked about.

Obviously not expecting any secrets to be spilled here, but getting some interesting thoughts on ancillary factors for modeling implied market sentiment beyond IV would be useful convo.



Using the past 12 months of price data for HH nat gas futures as an example would be cool I know the fin news talked about that foolhardy fellow who ran an 'options selling'/short-vol subscription service, but while that may well have exacerbated the upside vol given a decent aggregate position, that was just a side effect.

[Strange]

Regarding the "the 3D distribution of option greeks vs expiry & moneyness/strike". I don't know if "many traders do it", as it's a very hard problem. All you see is open interest and don't know who is going to be hedging delta and/or vega and who's just riding it naked. You can make some educated guesses, but that's about as good as it goes. It gets a bit easier close to expiration because you can combine the outstanding interest with some tick data analysis to have a guess if any name is pinning the strike, but even that is a difficult game to play.



Some people look at the attributed aggressive trades (i.e. use tick data for options to figure out if some strike was bought or sold aggressively and assume that customers don't hedge). Even if that assumption is correct (not always), this approach ignores any OTC or dealer-arranged flow. For example, even if you do know that a particular strike was a sell, there might be an exotic behind it that will totally negate your expectation of increased gamma hedges.

[JTDerp]

Thanks for the perspective, Strange. Does seem that, only close to expiration, would trying to game some reversionary-type of trade from the surface dynamics have a decent probability...otherwise it's just a (limited) looking glass and could be several larger themes at play in where things could go, i.e. what you said better than I